Heat, not light, is usually the limiting factor on a bright roof

It surprises people that solar panels are happiest in cold, bright weather rather than in a heatwave. Panels are rated in laboratory conditions at a cell temperature of 25°C, and every module datasheet carries a temperature coefficient: a figure describing how much power the panel loses for each degree its cells sit above that reference point. On a still British afternoon in July, cell temperatures can climb well above the air temperature, because the panel is absorbing energy it cannot convert and has only two ways to shed it — radiating it away, and handing it to moving air.

This is why a crisp, sunny day in March can feel oddly productive on a monitoring app while a sweltering August afternoon underperforms your expectations. The light is stronger in August; the cells are simply less efficient at converting it. Understanding that single relationship is enough to make the whole in-roof versus on-roof conversation obvious, because construction method decides how easily heat leaves the back of the array.

The air gap behind an on-roof array works like a chimney

A conventional on-roof installation fixes brackets through to the rafters, runs horizontal or vertical rails across the roof, and clamps the panels to those rails. That geometry leaves a cavity — typically a modest gap of a few centimetres, set by the mounting system and the manufacturer's installation manual — between the back of the glass-and-laminate sandwich and the tiles or slates beneath.

That cavity is not wasted space. It is a chimney. Cooler air is drawn in along the lower edge of the array, warms as it passes the hot rear surface of the panels, becomes less dense, rises, and exits along the top edge. The flow is gentle and largely invisible, but it runs all day whenever the panels are warmer than the surrounding air, and it is reinforced by any breeze crossing the roof plane. The result is a cell temperature meaningfully closer to ambient than a sealed installation would achieve.

Roof-integrated panels replace tiles rather than sitting above them. They are set into a tray or flashing system that forms the weatherproof layer itself, which is exactly why they look so clean — and exactly why the convection path is restricted. Some in-roof systems are designed with deliberate ventilation channels and perform better than others, but as a family they sit tighter, breathe less, and hold heat for longer.

When roof-integrated solar is still the right answer

Knowledge beats marketing in both directions, so it would be dishonest to present this as a simple win for on-roof mounting. There are situations where flush integration earns its small output cost several times over.

The clearest is a new build or a full re-roof. If the tiles are coming off anyway, an integrated system replaces roof covering rather than adding to it, which changes the material and labour arithmetic considerably compared with retrofitting the same panels onto a finished roof. Appearance is the second reason, and it is a legitimate one: on a prominent front elevation, a conservation-sensitive setting, or a house where the roofline is the main architectural feature, a flush array can be the difference between a system the household is happy to look at every day and one it quietly resents. Where planning sensitivities or a design-led scheme push strongly towards minimal profile, that matters.

There are also practical considerations that cut both ways. A flush array has no raised edges for wind to catch and no cavity for birds to nest in — a genuine nuisance on some on-roof installations, usually solved with mesh. Against that, an on-roof array keeps the existing weatherproof roof covering intact underneath it, which some homeowners prefer for long-term maintenance reasons. Neither is a trump card. They are inputs to a decision.

Other things that change how hot your array runs

Mounting method is the biggest single lever, but it is not the only one. Roof pitch and orientation affect how readily warm air moves up the cavity and how long the panels sit under peak irradiance. Array shape matters too: a tall block of panels has a longer chimney and a warmer top row than a wide, shallow single-row layout, because air entering at the bottom has already been heated by the time it reaches the upper modules.

Roof colour and construction have an influence, since a dark, heavily insulated roof re-radiates heat back at the panels more readily than a lighter, ventilated one. Local shelter counts as well — an array tucked behind a tall hedge or in an enclosed courtyard sees less passing air than one on an exposed gable. None of these individually rewrites a system's annual yield, but taken together they explain why two apparently identical installations can show slightly different summer behaviour, and why a proper survey looks at the roof rather than just counting panels.

Follow-up questions people usually ask next

Should I hose my panels down in a heatwave to cool them? No. The thermal benefit is brief, thermal shock on hot glass is an unnecessary risk, and climbing onto a roof to do it is a far greater hazard than a few lost kilowatt-hours.

Does the gap make the array noisier or draughtier indoors? The cavity sits outside the weatherproof roof covering, so it has no direct path into the loft. Wind noise around raised panel edges is occasionally noticed on very exposed properties, but it is not a common complaint.

Do I lose power on cloudy days too? You generate less, but UK homes still produce useful solar power under overcast skies — panels respond to diffuse light, not just direct sunbeams. Helpfully, those are also the conditions where cells run coolest and mounting method matters least. Does the same logic apply to ground-mounted arrays? Yes, and more favourably. A free-standing frame is ventilated on both faces, so it typically runs cooler than either roof option in equivalent weather. Will a battery change the calculation? Not the thermal physics, but it changes how much of your generation you actually use. If the question is overall system value, storage and consumption patterns usually move the needle more than a few summer percent. Can I mix both on one house? Yes — the episode shows exactly that, a rail-mounted section meeting a flush section on the same roof plane. It is sometimes the right answer where one elevation is prominent and another is not.

Video transcript

Mo: The flush ones look neater. Any cost? RoboMo: Sunlight becomes heat. Trapped behind flush panels, warmer cells produce less.

Mo: So the gap behind them is actually doing something? RoboMo: It is a chimney. Cool air enters low, warm air exits high, cells run cooler.

Mo: How much output does the flush side actually give up? RoboMo: A few percent across a year. Small, permanent, and paid for in looks. Mo: The flush ones look neater. Any cost? RoboMo: Sunlight becomes heat. Trapped behind flush panels, warmer cells produce less. / Mo: So the gap behind them is actually doing something? RoboMo: It is a chimney. Cool air enters low, warm air exits high, cells run cooler. / Mo: How much output does the flush side actually give up? RoboMo: A few percent across a year. Small, permanent, and paid for in looks.

Wrapping up

The gap behind an on-roof array is one of those quiet pieces of engineering that never appears in a brochure because it does not photograph well. It is just air, moving slowly, doing the useful work of carrying heat off the back of your panels so the cells stay closer to the temperature at which they perform best. Roof-integrated panels restrict that flow, run warmer under summer sun, and hand back a few percent of annual output in exchange for a roofline that sits flush and looks deliberate. That is a real cost and a real benefit, and which one matters more genuinely depends on your house, your roof, and what you want to look at. What should not happen is discovering the trade-off after the scaffolding has come down. Ask the question early, price both constructions on the same roof, and choose with your eyes open.